Comprehensive investigation of DNA damage repair genes in children with cancer identifies SMARCAL1 as novel osteosarcoma predisposition gene

Ninad Oak1, Wenan Chen2,3, Alise Blake1, Lynn Harrison1, Martha O’Brien4,5,6, Christopher Previti4,5,6, Gnanaprakash Balasubramanian4,5,6, Kendra Maass4,5,6,7, Steffen Hirsch4,8, Judith Penkert9,10, Barbara C Jones4,6,7,11, Kathrin Schramm4, Michaela Nathrath9,10, Kristian W Pajtler4,5,6,7, David T.W. Jones4,6,7,11, Olaf Witt4,6,7,12,13, Uta Dirksen14,15,16, Jiaming Li17, Yadav Sapkota18, Kirsten K Ness18, Lillian M Guenther1, Stefan M Pfister4,5,6,7, Christian Kratz9, Zhaoming Wang18,19, Greg T Armstrong18, Melissa M Hudson1,18, Gang Wu2,20, Robert J Autry4,5,6*, Kim E Nichols1*, Richa Sharma21*

* Corresponding authors

1Department of Oncology, St. Jude Children’s Research Hospital, Memphis, TN

2Center for Applied Bioinformatics, St. Jude Children’s Research Hospital, Memphis, TN

3Division of Computational Biology, Mayo Clinic, Rochester, MN

4Hopp Children’s Cancer Center Heidelberg (KiTZ), Heidelberg, Germany

5Division of Pediatric Neurooncology, German Cancer Research Center (DKFZ), Heidelberg, Germany

6German Cancer Consortium (DKTK), Heidelberg, Germany

7Department of Pediatric Oncology, Hematology and Immunology, Heidelberg University Hospital, Heidelberg, Germany

8Department of Human Genetics, Institute of Human Genetics, Heidelberg University Hospital, Heidelberg, Germany

9Pediatric Hematology and Oncology, Hannover Medical School, Hannover, Germany

10Department of Human Genetics, Hannover Medical School, Hannover, Germany

11Division of Pediatric Glioma Research, German Cancer Research Center (DKFZ), Heidelberg, Germany

medRxiv preprint doi: https://doi.org/10.1101/2025.05.12.25325832; this version posted June 4, 2025. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC 4.0 International license .

12National Center for Tumor Diseases (NCT) Heidelberg, Germany

13CCU Pediatric Oncol, German Cancer Research Center (DKFZ), Heidelberg, Germany

14Pediatrics III, West German Cancer Center University Hospital Essen, Essen, Germany 15German Cancer Consortium (DKTK) site Essen , Germany

16National Center for Tumor diseases (NCT)site Essen , Germany

17Department of Biostatistics, St. Jude Children’s Research Hospital, Memphis, TN 38105, USA

18Department of Epidemiology and Cancer Control, St. Jude Children’s Research Hospital, Memphis, TN

19 Department of Computational Biology, St. Jude Children’s Research Hospital, Memphis, TN

20Department of Pathology, St. Jude Children’s Research Hospital, Memphis, TN

21 Department of Hematology, St. Jude Children’s Research Hospital, Memphis, TN

Corresponding authors:

Richa Sharma, MD (Current Address) Cleveland Clinic Children’s 9500 Euclid Avenue, Cleveland, OH 44195 Email: sharmar19@ccf.org Phone: (317) 658-5120

Kim E. Nichols, MD St. Jude Children’s Research Hospital 262 Danny Thomas Place, MS 1170, Room I3311, Memphis, TN 38105-3678 Email: Kim.Nichols@stjude.org Phone: (901) 595-8385

Robert J Autry, PhD

KiTZ - Hopp Children’s Cancer Center Heidelberg Im Neuenheimer Feld 580, D- 69120 Heidelberg Germany Email: robert.autry@kitz-heidelberg.de

Phone: (901) 233-3922; (+49) 178 261-8801

Running Title: SMARCAL1 germline variation and pediatric osteosarcoma

Conflict of Interest

No author disclosures were reported.

Abstract

Background: Recent large-scale genomic sequencing studies reveal that 5-18% of children with cancer harbor pathogenic variants (PV) in known cancer predisposing genes (CPG). However, DNA damage repair (DDR) genes, which are frequently somatically altered in pediatric tumors, have not been systematically examined as a source of novel cancer predisposing signals.

Methods: To address this gap, we interrogated 189 genes across six DDR pathways for the presence of PV among 5,993 childhood cancer cases and 14,477 adult non-cancer controls. PV were defined as rare (allele frequency <0.05% in the gnomAD v2.1 non-cancer subset), nonsense, frameshift, affecting canonical splice sites, and missense with REVEL score >0.7. Using logistic and firth regression, we identified genes with statistically enriched PV and replicated findings among 1,494 additional childhood cancer cases across three independent cohorts.

Findings: Analysis across all cancers revealed enrichment of TP53 PV (0.6%, false discovery rate [FDR]logistic=0.0066, FDRFirth=0.0064). Cancer-specific analyses confirmed previously identified associations for germline TP53 PV in adrenocortical carcinoma (37%, FDRlogistic<0.0001, FDRFirth=0) and high-grade glioma (2.4%, FDRlogistic=0.0022, FDRFirth=0.1082), as well as BARD1 PV in neuroblastoma (1.2%, FDRlogistic=0.0341, FDRFirth=0.2682). Three novel gene-tumor associations were identified, including POLL PV in Ewing sarcoma (1.7%, FDRlogistic=0.0319, FDRFirth=0.3101), SMC5 PV in medulloblastoma (1.6%, FDRlogistic=0.0005, FDR Firth=0.0499) and SMARCAL1 PV in osteosarcoma (2.6%, FDRlogistic=0.0250, FDRFirth=0.2180). Among these putative CPG, SMARCAL1 PV were enriched in osteosarcoma across each of the replication pediatric cancer cohorts (2.5%, PFisher <0.0001). All three osteosarcomas with available tumor data exhibited deletion of the wild-type SMARCAL1 allele.

Interpretation: Our study identifies SMARCAL1 PV as a predisposing factor for osteosarcoma, providing insights into tumor biology and creating opportunities for development of novel therapeutic, surveillance, and preventive interventions for this aggressive childhood cancer.

Background

As the initiating genetic events in tumorigenesis, germline variants in cancer predisposing genes (CPG) perturb cell growth and differentiation to set the stage for malignant transformation. Accordingly, the study of CPG and associated hereditary syndromes provides critical insights into normal physiology and cancer biology. To this end, recent sequencing studies have revealed that up to 18% of children with cancer harbor an underlying genetic predisposition.1-4 However, 40-80% of children with cancer have family histories and/or clinical features concerning for cancer predisposition but lack a causal genetic diagnosis.1,5 This observation suggests that additional CPG remain to be discovered and their association with cancer phenotypes further elucidated.

Prior investigations have primarily included children with specific tumor types (e.g., high risk solid or central nervous system [CNS] tumors, relapsed cancers) and examined for pathogenic variants (PV) in known CPG. Nevertheless, expanding the scope of germline analyses to include children with a broader array of cancers and additional cancer-associated genes is crucial to identify the missing heritable factors underlying childhood tumor formation. The identification of novel CPG and predisposing variants is also central to improving the outcomes for affected children as it enables development of targeted cancer therapies, genetic counseling and testing of relatives, and improves approaches to cancer surveillance and risk reduction.62

Somatic alterations impacting DNA damage repair (DDR) genes are prominent drivers of high grade pediatric tumors.2 In addition, germline PV impacting selected DDR genes have been

linked to several highly penetrant childhood cancer predisposition syndromes (CPS), including Li-Fraumeni syndrome, ataxia telangiectasia, Fanconi anemia, and replication repair deficiency.2.8 Germline PV in DDR pathway genes have also been implicated in development of subsequent malignant neoplasms in long-term survivors of childhood cancer, especially those previously exposed to higher doses of ionizing radiation, anthracyclines or alkylating agents.º Despite these prior observations and to the best of our knowledge, an unbiased assessment of DDR pathway genes and their role in development of primary cancers in children has not been conducted.

To this end, we generated a harmonized dataset of germline variants from 5,993 childhood cancer cases and 14,477 adult non-cancer controls. We then conducted rare variant gene burden analysis using a curated set of 189 DDR genes with the aim to identify novel CPG that could account for the missing heritability of childhood cancer. Novel gene-cancer associations were replicated using three independent pediatric cancer cohorts and available tumor data.

Methods

Patient Cohorts

The discovery cohort consisted of 5,993 children with cancer across five large scale sequencing studies, including the Pediatric Cancer Genome Project (PCGP, phs000409),1º National Cancer Institute Therapeutically Applicable Research to Generate Effective Treatments initiative (NCI- TARGET, phs000218),1 St. Jude Lifetime Cohort Study (SJLIFE),12 Genomes for Kids (G4K),3 and St. Jude Real-Time Clinical Genomics study (RTCG) (Figure 1A, Supplementary Table S1). The control cohort for discovery comprised 14,477 adults without cancer from the 1000 Genomes Project13 and Alzheimer’s Disease Sequencing Project (phs000572).14 This study was approved by the Institutional Review Board at St. Jude Children’s Research Hospital (No. 20- 0379). For replication of novel CPG, we queried three independent pediatric cancer cohorts, including the Childhood Cancer Survivorship Study (CCSS, phs001327),15 INdividualized Therapy FOr Relapsed Malignancies in Childhood (INFORM)16 and the German Childhood Cancer Registry (GCCR). The control cohort for replication included adults without cancer from gnomAD v2.117.

Variant Calling and Filtering for the Discovery cohort

Germline whole exome sequencing (WES) reads from cases and controls were mapped to GRCh37 with Burrow-Wheeler Aligner followed by joint variant calling using GATK best practices workflow with modifications (Supplementary Methods). Downstream analyses were restricted to germline variants in 189 genes that function in DDR (Supplementary Table S2).89 We filtered to retain rare damaging (i.e., pathogenic) variants (PV) if the they fulfilled the following criteria: 1) minor allele frequency (AF) <0.05% in gnomAD v2.1 (non-cancer subset);17 2) nonsense or frameshift variants; 3) missense variants with REVEL scores >0.7; and 4) canonical splice site variants. Matched tumor data for specific samples in our discovery cohort were analyzed as described in the Supplementary Methods.

Statistical Analysis

We performed a gene-based burden analysis for rare damaging germline variants in DDR genes in the discovery pediatric cancer cohort versus non-cancer controls. Twenty-two cancer types with at least 25 cases were considered for this analysis. For gene burden analysis, we applied logistic regression (glm-logit, R package stats v3.6.2; hereafter represented as false discovery rate [FDR]logistic) and Firth regression (R package logistf v1.26; represented as FDRFirth with odds ratio [OR]) using the first five principal components as co-variates to account for genetic ancestry. We computed nominal p-values using the Chi-squared and Walds test, followed by correction for multiple comparisons using the Benjamini & Hochberg method. Significance was determined as FDR <0.05. For replication of statistically significant gene- cancer associations, we queried relevant pediatric cancer cases in CCSS, INFORM, GCCR, and adult non-cancer controls (gnomAD v2.1) followed by filtering germline PV based on criteria similar to the discovery cohort. Fisher exact test was used to evaluate for enrichment of damaging germline variants in respective cancer types. Comparison of age at cancer diagnosis between DDR germline variant carriers and non-carriers was completed as described in the Supplementary Methods.

Results

Germline variants in DDR genes across tumor types

The discovery cohort included 5,993 children and adolescents with 22 cancer subtypes, broadly classified as hematologic, solid, and CNS tumors (Figure 1A, B). The median age at cancer diagnosis was 6 years (4 days to 32 years) with most cases (67%) of European ancestry (Supplementary Table S3). Among the 6,636,054 germline variants filtered from cases, 2,059 rare, protein truncating, splicing, or predicted damaging missense variants among 189 DDR genes were retained (Figure 1C, Supplementary Table S4). The prevalence of putative PV among the 189 DDR genes in cases was comparable to controls (26.7%; p=0.07; not shown). Additionally, the frequency of PV was similar across hematologic (mean prevalence 29.9% [range 26.9% - 33.7%]), solid (31.1% [20.7% - 55.6%]), and CNS tumors (28.7% [23.3% - 36.4%]; p=0.8, Kruskal-Wallis test) but varied across cancer subtypes, with the lowest prevalence in retinoblastoma (20.7%) and highest in adrenocortical carcinoma (ACC, 55.6%) (Figure 1D).

Significant germline associations in DDR genes

Overall, 1,697 of 5,993 (28.3%) childhood cancer cases harbored putative PV in one or more DDR genes. Analyses across all cancers revealed significant enrichment of PV in TP53 compared to jointly-called adult non-cancer controls (n=14,477) (0.6%, FDRlogistic=0.0066; FDRFirth=0.0064, OR=2.8 [95% CI: 1.7 - 4.6]), supporting its critical role in maintaining genome stability (Supplementary Table S5). We next tested for enrichment of damaging germline variants in DDR genes across the 22 cancer subtypes in our discovery cohort. We observed enrichment of TP53 PV in ten of 27 ACC cases (37%, FDRlogistic<0.0001; FDRFirth=0, OR=289.2 [120.7 - 673.5]) and five of 206 high-grade glioma (HGG) cases (2.4%, FDRlogistic=0.0022; FDR Firth=0.1082, OR=11.1 [3.9 - 26.0]) (Figure 1E, Table 1, Supplementary Table S5, S6). Both germline TP53-tumor associations have previously been described and serve as internal

validation of our analytic pipeline.1.18 Of note, the founder variant, TP53:p.R337H, common in southern Brazil, did not meet our filtering criteria (REVEL=0.693) and therefore, was excluded from this analysis. Among the 14 unique germline PV in TP53, six were protein truncating or splicing variants and the remaining were missense variants reported to negatively impact transcriptional activity (Figure 2A).19,20 Tumor WES or WGS data were available for each of the germline TP53-mutated ACC and HGG, which showed a somatic second hit affecting the remaining TP53 allele in all cases (Supplementary Figure S1A).

The BARD1 gene, which encodes BRCA1-associated RING domain 1, has emerged as a neuroblastoma (NBL) predisposition gene.21 Similar to prior reports, we observed germline BARD1 variants in six of 485 NBL cases (1.2%; FDRlogistic=0.0341; FDRFirth=0.2682, OR=6.0 [2.3 - 13.2]), including four predicted as truncating and two that were missense (Figure 2A, Supplementary Tables S5, S6). The missense variants (p.L480S, p.L447V) are located in the ankyrin repeat domain of BARD1, a region in which mutations cause HR deficiency (Figure 2B). 22 Tumor WES data were available for all six NBL cases and WGS data for two cases. Second somatic hits affecting BARD1 were not observed in any of these cases (Supplementary Figure S1A). In the two tumors with WGS data, DNA mutation signature analysis revealed presence of SBS18, associated with reactive oxygen species damage, a signature observed in NBL (Figure 2C).23 RNAseq data were available for three tumors, one of which (germline p.Y180*) showed BARD1 expression at less than 10th percentile compared to non-BARD1 mutated NBL tumors (Supplementary Figure S1B). While this finding suggests loss-of-heterozygosity (LOH) in the tumor, tumor WGS data were not available to corroborate this possibility.

In addition to verifying known associations, we identified three novel gene-cancer associations, including POLL in Ewing sarcoma (EWS), SMC5 in medulloblastoma (MB), and SMARCAL1 in

osteosarcoma (OS) (Table 1, Supplementary Table S5). Two of 117 (1.7%; FDRlogistic=0.0319; FDRFirth=0.3101, OR=23.8 [4.6 - 81.1]) EWS cases exhibited germline damaging variants in POLL (Figure 2A, Supplementary Table S6). POLL encodes the DNA polymerase lambda, which performs 3’-end extension during non-homologous end joining (NHEJ) and base excision repair (BER). Both p.Q469* and p.R487L are located in the nucleotidyltransferase domain within the DNA binding groove of POLL that is important for gap filling during NHEJ (Figure 2B). Neither variant is included in ClinVar and the missense variant p.R487L is predicted as LP by AlphaMissense. Tumor WES and RNAseq data were available for the p.Q469 *- associated tumor, which revealed the EWS somatic driver, EWSR1:FLI1, and no additional SNVs in POLL (Supplementary Figure S1A, Supplementary Table S6), which is consistent with heterozygosity in the tumor and supported by the observed 63rd percentile POLL expression noted by tumor RNAseq (Supplementary Figure S1B).

Four of 257 MB cases (1.6%; FDRlogistic=0.0005; FDRFirth=0.0499, OR=21.6 [6.4 - 60.1]) exhibited germline PV in SMC5, the gene encoding Structural Maintenance of Chromosome 5 (Figure 2A). SMC5 is important for DNA replication, repair and chromosome maintenance with biallelic germline alterations associated with the neurodevelopmental disorder, Atelis syndrome- 2.24 Interestingly, all four germline SMC5 mutant cases were of the group 3 MB subtype. The association of germline SMC5 PV with MB was even stronger when analyses were restricted to include only group 3/4 MB (four of 88 cases, 4.5%; FDRlogistic<0.0001; FDRFirth=0.0011, OR=54.6 [16.0 - 156.0]) (Supplementary Table S5). The p.E126Vfs*12 and c.380+1G>C, located in the N-terminus P-loop NTPase domain, and the p.C417* variant, located in the coiled-coil domain, are predicted to result in nonsense-mediated decay (NMD). The missense variant p.N940Y was located in the C- terminal P-loop NTPase domain of SMC5, important for DNA binding and ATP- driven loop extrusion by the SMC5/6 complex (Figure 2A-B). Although none of the SMC5 variants were reported in ClinVar, p.N940Y is predicted to be LP by AlphaMissense. Tumor

RNAseq and WES data were available for all four SMC5 germline variant carriers, whereas WGS was available for three. Tumor SMC5 expression was variable across the four cases. The MB tumor associated with germline c.380+1G>C, resulting in an out-of-frame transcript, showed exon 3 skipping in ~50% of SMC5 transcripts (Figure 2C) and SMC5 expression at <10th percentile compared to SMC5 wild-type MB (Supplementary Figure S1B). Due to lack of tumor WGS data for this case, we could not determine whether there was a deletion affecting the remaining SMC5 allele. We identified a second somatic hit, SMC5:p.Q357K, in the p.N940Y germline-mutated MB but could not establish cis versus trans allelic configuration of the two variants. Of note, this tumor exhibited SBS8, a DNA mutational signature associated with late replication errors in cancer (Figure 2C), supporting SMC5’s role in mitotic progression.25

Finally, we identified six of 230 OS cases (2.6%; FDRlogistic=0.0250; FDRFirth=0.2180, OR=6.26 [2.4 - 13.5]) with germline SMARCAL1 variants (Figure 3A, Supplementary Tables S5, S6). SMARCAL1 encodes the SNF2 related Chromatin Remodeling Annealing Helicase, which resolves stalled replication forks and secondary DNA structures to support DNA replication and repair.26 We identified four protein-truncating variants (p.R114Qfs*4, p.L139Efs*3, p.L397Rfs*40, p.Q653*) that are predicted to undergo NMD and two missense variants (p.R820H, p.R490C) that are located in the SMARCAL1 helicase domain. The p.R820H variant was classified as P by ClinVar and predicted as LP by AlphaMissense, while p.R490C was classified as a VUS in ClinVar and ambiguous by AlphaMissense (Figure 3B-C, Supplementary Table S6). Considering that OS is commonly observed in Li-Fraumeni syndrome, we queried and confirmed the absence of germline TP53 mutations in the six germline SMARCAL1-mutated cases. Further, none of the OS cases harbored germline variants in a list of 60 known CPG associated with autosomal dominant CPS of moderate to high penetrance.27 Matched tumor WES and RNAseq data were available for two cases, which did not reveal a somatic second hit (SNV) in SMARCAL1. The germline p.R114Qfs*4-mutated OS harbored a somatic

ATRX:p.P717Hfs*4 mutation. Tumor RNAseq from this case showed unaffected SMARCAL1 expression, while the germline p.Q653 *- mutated case demonstrated reduced SMARCAL1 RNA expression <1st percentile. However, additional genomic alterations associated with this reduced expression could not be ascertained in the absence of tumor WGS data (Figure 3B, Supplementary Table S6, Supplementary Figure S1B).

Clinical features associated with germline DDR gene variants identified in the Discovery cohort Patients with germline CPG PV are often younger at tumor onset than individuals with sporadic cancers. Therefore, we analyzed the ages of cancer onset in carriers versus non-carriers of damaging DDR gene variants for the significant gene-cancer associations. Through this analysis, we observed a younger age of onset of 20 months in ACC cases carrying TP53 PV versus 60 months (Wilcoxon rank-sum test p<0.05) in non-carriers (Supplementary Table S7). We did not find significant differences in ages of cancer onset for any of the other significant gene-cancer associations, or any DDR PV carriers across all cancers when compared to non- carriers. Family history information was available for 17 of 33 (52%) patients with significant gene-cancer associations, eight of whom had a positive family history of cancer, defined as having ≥1 first- or second-degree relatives under 50 years of age with cancer or tumor (excluding cervical and non-melanoma skin cancer; Supplementary Table S6). However, none of the relatives from cases with a positive family cancer history had tumors similar to the cases described here.

Replication of SMARCAL1 as a novel osteosarcoma predisposition gene

To replicate our novel gene-tumor associations, we queried three additional pediatric cancer cohorts (CCSS, GCCR, and INFORM). POLL did not replicate and SMC5 reached significance in an inadequately powered GCCR MB cohort (one of 30; 3.3%; PFisher=0.020, OR= 52 [1.3 - 309.1]. Importantly, we confirmed significant enrichment of damaging SMARCAL1 variants in 15

children with OS across CCSS (eight of 272 cases; 2.9%; Fisher exact PFisher<0.0001, OR= 8.6 [3.7 - 17.3]), GCCR (four of 135 cases; 3%; PFisher=0.001, OR= 8.7 [2.3 - 22.6]), and INFORM (three of 197 cases; 1.5%; PFisher =0.032, OR= 4.4 [0.9 - 13.1]) compared to 0.35% (409 in 118,184) in adult non-cancer controls (gnomAD v2.1) (Supplementary Table S8). These 15 cases harbored 12 unique germline variants, two of which were also observed in the discovery cohort (p.L139Efs*3, p.L397Rfs*40). Among the 10 remaining SMARCAL1 variants, three were protein-truncating (p.F941Lfs*31, p.R563*, p.E848*), three canonical splice-site (c.863-2A>G, c. 1335-2A>T, c.2070+2dup) and four missense (p.F801V, p.A838T, p.G857R, p.F279S). Three of the missense variants (p.F801V, p.A838T, p.G857R) are clustered within the SMARCAL1 helicase domain that is critical for DNA binding, while one (p.F279S) is located in the Hep-A- related protein (HARP) domain, essential for replication fork stabilization (Figure 3C).26 Five of these 10 variants are reported as P or LP and four as VUS in ClinVar, while one is unreported (Supplementary Table S9). Three missense variants (p.F801V, p.G857R, p.F279S) are predicted to be LP by AlphaMissense, suggesting an adverse impact on SMARCAL1 structure (Figure 3C, Supplementary Table S9). None of the 15 SMARCAL1 germline variant carriers from these replication cohorts harbored a PV in TP53; however, two cases harbored PV affecting other CPG: one in NF2 (c.243-2A>C, CCSS) and one in MLH1 (p.615_616del, INFORM) (Figure 3B).2 In sum, we identified 12 unique damaging germline SMARCAL1 variants in 15 OS cases across the replication cohorts for a prevalence of 2.5% (PFisher <0.0001, OR= 7.3 [4.0 - 12.2]), which is similar to the 2.6% prevalence in the discovery cohort (Table 1, Figure 3A, B; Supplementary Table S8, S9).

Across replication cohorts, tumor WGS data were available only for the INFORM cases. We observed SMARCAL1 LOH in all three relapsed OS tumors due to copy number variation (Supplementary Figure S2). Two of three tumors also had RNAseq data available, one of which (c.1335-2A>T) exhibited low SMARCAL1 expression (Supplementary Figure S3). Loss

of SMARCAL1 function has been shown to result in alternative-lengthening of telomeres (ALT) in glioblastoma cells.26 To this end, we observed all three OS tumors in the replication cohort to be ALT positive as determined by TelomerHunter28 (Supplementary Material). In addition, we assessed tumor genomic data in these three cases for variations in ALT associated genes, including ATRX, DAXX and H3F3A and did not observe somatic or germline SNVs (n=3 WES) or reduction in RNA expression (n=2 RNAseq) in these genes. These data suggest that loss of SMARCAL1 function may induce ALT, a potential mechanism by which it promotes osteosarcoma formation (Figure 3B, Supplementary Figure S3).

Discussion

Highly penetrant CPS result from germline variations in DNA damage repair (DDR) genes. To date, comprehensive studies investigating the germline landscape of DDR alterations in pediatric cancers have been lacking, in part due to the lack of case-control designs that harmonize differences such as batch effects from library preparation, sequencing platforms, and variant calling pipelines necessary for identifying novel cancer-predisposing variants. To circumvent these issues, in this study we re-mapped raw sequencing data and performed joint genotype calling across 5,993 pediatric primary cancer cases and 14,477 adult non-cancer controls. Further, we only included variants based on stringent criteria, including rarity in the general population and presumed negative impact protein expression or function based on in- silico pathogenicity predictions and prior interpretations available through established variant curation databases such as ClinVar. Through this approach, we established a 28.3% prevalence of putative damaging variants in DDR genes across a wide array of childhood cancers. Statistical testing confirmed known associations, including germline TP53 PV in ACC and HGG, and BARD1 PV in NBL. Importantly, we discovered novel associations of POLL PV in EWS, SMC5 PV in MB, and SMARCAL1 PV in OS. Using three independent replication cohorts, we confirmed statistical enrichment of germline damaging variants in SMARCAL1, highlighting it as a novel predisposing gene in which damaging variants increase the risk for pediatric OS.

We find that 2.5% of OS cases harbor germline SMARCAL1 variants of which two-thirds are expected to cause haploinsufficiency due to protein truncation while one-third are missense and predicted to be damaging. Ballinger et al., previously reported germline truncating SMARCAL1 variants in 19 sarcoma cases including two OS, 29 while Akhavanfard et al., who used data from the SJLIFE cohort, identified SMARCAL1 in three OS cases.3º Moreover, two cases of OS have been reported in individuals with Schimke immune-osseous dysplasia (SIOD), a multisystemic disorder secondary to biallelic SMARCAL1 loss-of-function alterations that is typified by bony

defects (e.g., short stature, bony dysplasias) suggesting that OS is a cancer associated with SMARCAL1 dysfunction.31 However, the rarity and short life span (median age of death: 11years) of individuals with SIOD preclude us from accurately assessing the prevalence and penetrance of OS in this context. No other germline alterations in OS-relevant CPG were identified in our cases, which supports SMARCAL1 as an independent risk factor for OS.32 Six of 12 osteosarcoma cases (50%) with available clinical information experienced disease relapse. Longitudinal studies are needed to define the penetrance and clinical features characteristic of OS associated with germline SMARCAL1 variation.

Somatic LOF mutations in SMARCAL1 have been identified in glioblastoma resulting in SMARCAL1 deficiency and permitting ALT-mediated telomere synthesis, a homologous recombination-based mechanism of telomere elongation utilized by 10-15% of high-risk cancers.26,33 Consistent with these findings, ALT appeared active in all three tumors from INFORM with bi-allelic SMARCAL1 alterations. ALT-positive tumors also frequently harbor LOF mutations in chromatin modification genes including, ATRX, DAXX and H3F3A, which contribute to ALT induction by dysregulating histone H3.3 deposition at telomeres.34 While we did not observe alterations in these genes in the three tumors with bi-allelic SMARCAL1 mutations, one of two tumors with available data in our discovery cohort harbored an inactivating ATRX mutation. Ballinger et al., also reported LOH in 5 of 19 sarcomas with germline SMARCAL1 variants; however, it is not possible to discern whether these were OS cases.29 Altogether, these data suggest that SMARCAL1 functions as a tumor suppressor. We propose a model whereby germline SMARCAL1 PV cause partial or altered SMARCAL1 function, leading to DNA replication and repair defects throughout the genome, including the telomeres, with acquisition of second somatic hits in ALT permissive genes, which results in OS tumor formation (Figure 3D).

In addition to SMARCAL1, we report a novel association of SMC5 in group 3 MB and POLL in EWS. Out of four individuals with germline SMC5 PV, one harbored a second somatic hit in SMC5, resulting in a late replication error mutational signature, which may correlate with this patient’s aggressive disease features (metastatic disease at diagnosis, relapse, death). Future studies in larger MB cohorts are required to validate this association and the role of SMC5 dysfunction in high grade MB. Limited data are available for the role of the polymerase lambda gene, POLL, in cancers, including EWS. Loss of POLL leads to BER deficiency in mouse embryonic fibroblasts.35 Tumors harboring the EWS driver, EWSR1:FLI1, which was observed in one of two POLL germline mutated EWS tumors, are significantly sensitive to the PARP inhibitor olaparib, which is further accentuated in the presence of other DNA damaging agents. 36 Although clinical response to PARP inhibitors in EWS is underwhelming, cases where the BER pathway is impaired, such as those with POLL-mediated predisposition, may constitute a subset with improved outcomes.

The following limitations should be considered when interpreting the results of this study. First, to achieve adequate statistical power, we combined high-risk primary cancers with presumably lower risk cancers from adult survivors of childhood cancer. This approach may have diluted genetic signals in primary malignancies and added survival bias when examining the risk of childhood cancer, which warrants reassessment of primary cancers as larger cohorts become available. In our study, the variant filtering criteria were rigorous, and it is possible that additional clinically relevant germline variants with lower REVEL scores remain undescribed. To test this in our INFORM replication cohort, we decreased the filtering REVEL score to 0.5 and identified one germline SMARCAL1 missense variant (p.D424V) in two unrelated OS cases. Matched tumor data from these cases revealed somatic second hits in the remaining wild-type SMARCAL1 allele; however, we have not included these cases in our statistical analyses to maintain consistency in how the discovery and replication cohorts were analyzed. Future

studies are warranted to determine the extent to which these and other germline SMARCAL1 variants play a role in OS tumorigenesis. Notably, associations of SMARCAL1 with OS and SMC5 with MB remain statistically enriched even when analyses are restricted to include only protein-truncating and splicing variants (Supplementary Table S5). Finally, we could not establish the mode of inheritance or co-segregation with other cancers for many of the identified variants due lack of familial testing. Future efforts examining the relatives of germline SMARCAL1 PV carriers are needed as this information may serve to strengthen evidence of pathogenicity. Finally, assessment of tumor genomic data is critical for determining the functional consequence of predisposing variants; however, these data were available for only a limited number of cases.

In summary, outcomes for children with OS, especially those with relapsed or metastatic disease, remain suboptimal. Our finding that germline mutations in SMARCAL1 predispose to OS serves as a foundation for future studies aimed at developing novel therapies for this aggressive cancer, one for which there have been little advancements in treatment over the last four decades.37 Similarly, genetic testing for germline SMARCAL1 PV will enable prospective surveillance for the detection and treatment of incipient OS tumors at their earliest and most curable stages.

Acknowledgements

Funding for this study was provided by the American Lebanese Syrian associated charities. This study was supported by the following National Cancer Institute grants, R01CA283333 (Zhaoming Wang and Kim E Nichols), The St. Jude Lifetime Cohort (SJLIFE) (CA195547, M.M. Hudson, K.K. Ness), and The Childhood Cancer Survivor Study (CCSS) (CA55727, G.T. Armstrong). The INFORM program is financially supported by the German Cancer Research Center (DKFZ), several German health insurance companies, the German Cancer Consortium (DKTK), the German Federal Ministry of Education and Research (BMBF), the German Federal Ministry of Health (BMG), the Ministry of Science, Research and the Arts of the State of Baden- Württemberg (MWK BW); the German Cancer Aid (DKH), the German Childhood Cancer Foundation (DKS), RTL television, the aid organization BILD hilft e.V. (Ein Herz für Kinder) and the generous private donation of the Scheu family. We would like to express our sincere thanks to Carsten Maus, Erjia Wang (Next Generation Sequencing Core Facility, DKFZ). Lena Weiser, Gregor Warsow (Omics IT and Data Management Core Facility, DKFZ) for their highly dedicated support in data management and processing and Rolf Kabbe (Division of Pediatric Neurooncology, DKFZ) for his sincere and dedicated contribution to the bioinformatics analyses. Biostatistics support is provided by the Biostatistics Shared Resource (BSR) of the St. Jude Children’s Research Hospital and St. Jude Comprehensive Cancer Center (NIH P30CA021765). The authors thank the patients and families included in this study and members of the St. Jude Clinical Genomics Laboratory, without whom this work would not have been possible.

Data Availability Statement

The processed genomic data generated in this study are provided in the Supplementary Tables. Controlled access raw genomic data can be requested via St. Jude Cloud at https://platform.stjude.cloud/. The Childhood Cancer Survivor Study is a US National Cancer Institute funded resource (U24 CA55727) to promote and facilitate research among long-term survivors of cancer diagnosed during childhood and adolescence. CCSS data are publicly available on the St Jude Survivorship Portal within the St. Jude Cloud at https://survivorship.stjude.cloud/. In addition, use of the CCSS data that leverages the expertise of CCSS Statistical and Survivorship research and resources will be considered on a case-by case basis. For this use, a research Application of Intent followed by an Analysis Concept Proposal must be submitted for evaluation by the CCSS Publications Committee. Users interested accessing this resource are encouraged to visit http://ccss.stjude.org. Full analytical data sets associated with CCSS publications since January 2023 are available on the St. Jude

Survivorship	Portal	at			https://viz.stjude.cloud/community/cancer-survivorship-
community~4/publications.		Any	additional data	are	available	upon	request	from	the
corresponding author.

Figure Legends

Figure 1 Germline DDR gene variants across pediatric cancers.

(A) Study design and workflow. The discovery cohort comprised 5,993 pediatric cancer patients from the PCGP, G4K, RTCG, SJLIFE, and TARGET cohorts, with 14,477 controls from the 1000 Genomes Project and Alzheimer’s Disease Sequencing Project (ADSP). Germline variants were called using GATK joint genotyping (n=20,470) and underwent quality control and post hoc filtering along with genetic ancestry and sex determination. Variant filtering focused on 189 DNA repair genes, selecting rare variants (allele frequency <0.05%) predicted to be deleterious, including protein-truncating, splicing, and high-confidence missense variants (REVEL >0.7). Replication was performed using independent cohorts, including CCSS, GCCR, and INFORM. (B) Number of jointly called germline whole-exome sequencing samples across 22 pediatric cancers in the discovery cohort.

(C) Variant filtering pipeline. Among 5,993 pediatric cancer samples, 6,636,054 variants passed quality control. Filtering by predicted functional impact (frameshift, nonsense, missense, and splicing) yielded 2,536,772 variants, of which 1,403,469 were rare (gnomAD v2.1 non-cancer popmax AF <0.05%). Restriction to DNA damage repair (DDR) genes identified 10,368 variants, with 2,059 meeting the pathogenicity threshold (protein truncating, splicing, or damaging missense- REVEL >0.7).

(D) Prevalence of germline DDR gene variants across cases with 22 pediatric cancer types is shown. Median prevalence across the three major cancer classes (hematologic: red, CNS: green, solid: blue) is shown with a dashed line.

(E) Gene-level burden of pathogenic germline variants. All associations with FDR <0.25 are shown, whereas those with FDR <0.05 are highlighted with a black border. The color intensity represents -log10[FDR]). The number of variants for respective gene-cancer pairs is indicated. Abbreviations: ACC= Adrenocortical Carcinoma; ACPG= Adamantinomatous Type Craniopharyngioma; ALL-NOS= Acute Lymphoblastic Leukemia- Not Otherwise Specified;

AML= Acute Myeloid Leukemia; ATRT= Atypical Teratoid/Rhabdoid Tumor; BALL= B-Cell Acute Lymphoblastic Leukemia; CCSS= Childhood Cancer Survivorship Study; EPD= Ependymoma; EWS= Ewing Sarcoma; FDR = false discovery rate; G4K= Genomes4Kids; GCCR= German Childhood Cancer Registry; GCT= Germ Cell Tumor; HGG= High-Grade Glioma; HL= Hodgkin Lymphoma; INFORM= INdividualized Therapy FOr Relapsed Malignancies in Childhood; LGG= Low-Grade Glioma; LIC= Liver Cancer; MB= Medulloblastoma; MEL= Melanoma; NBL= Neuroblastoma; NHL= Non-Hodgkin Lymphoma; OS= Osteosarcoma; PCGP = Pediatric Cancer Genome Project; RB= Retinoblastoma; RMS= Rhabdomyosarcoma; RTCG= (St.Jude’s) Real- time Clinical Genomics; SJLIFE= St. Jude Lifetime Cohort; TALL= T-Cell Acute Lymphoblastic Leukemia; WLM= Wilms Tumor.

Figure 2: Germline mutations in DDR genes and their putative functional impacts.

(A) Distribution of predisposing variants along the protein sequences of TP53, BARD1, POLL, and SMC5 across adrenocortical carcinoma (ACC), high-grade glioma (HGG), neuroblastoma (NBL), Ewing sarcoma (EWS), and medulloblastoma (MB), respectively. Variants are categorized by type: frameshift (red), nonsense (orange), splice-site (purple), and missense (blue).

(B) Structural mapping of selected missense variants in BARD1, POLL, and SMC5 as predicted by AlphaFold. Germline variants such as p.L480S and p.L557V in the ankyrin repeat domain of BARD1, p.R487L in nucleotidyltransferase domain of POLL and p.N940Y in P-loop NTPase domain of SMC5 are located in regions critical for protein stability and/or function. * denotes variants predicted as P/LP by AlphaMissense.

(C) Somatic DNA mutational signature analysis across five cases with whole genome sequencing data was performed using signature.tools.db. Bar plots display the proportion of COSMIC (v3) single-base substitution (SBS) mutational signatures, including SBS1 and SBS5

(aging-related), SBS8 (late replication errors), and SBS18 (reactive oxygen species damage). Unassigned mutational signatures are also shown.

(D) Tumor RNAseq from the medulloblastoma (MB) case with the germline SMC5:c.380+1G>C splicing variant. The top panel (blue) represents normal splicing in another MB case without an SMC5 alteration, while the bottom panel (red) shows aberrant splicing in the SMC5:c.380+1G>C germline variant positive case. The variant leads to skipping of exon 3, as indicated by disrupted exon-exon junctions and altered splicing events, suggesting a pathogenic impact of the mutation.

Figure 3: Characterization of germline SMARCAL1 variant-positive osteosarcoma (OS). (A) Schematic representation of SMARCAL1 (NM_014140) protein with predisposing variants from the discovery cohorts (top) and replication cohorts (bottom). Protein domains RBD- RPA binding domain (green), HARP- HepA-related protein (blue) and Helicase (red) are shown. Variants are categorized by mutation type with frameshift (red), nonsense (orange), splice-site (purple), and missense variants (blue).

(B) Heatmap representation of SMARCAL1 variants in discovery and replication cohorts. Data include pathogenicity predictions based on ClinVar classification, REVEL scores, AlphaMissense predictions, and gnomADv2.1 allele frequencies. Additional pathogenic germline variants in SJCPG60 genes or somatic driver events are shown along with annotations for biallelic status (1), tumor relapse status and tumor availability.

(C) SMARCAL1 protein structure predicted by AlphaFold. Missense variants in discovery (green) and replication (purple) cohorts of OS are shown. Protein domains HARP- HepA-related protein (blue) and Helicase (red) are shown. * Denotes variants predicted as LP by AlphaMissense.

(D) Proposed model of SMARCAL1-mediated OS predisposition. SMARCAL1 is important for accurate DNA replication and repair to reinforce genome integrity. Approximately, 2.5% of

osteosarcoma cases carry a predisposing variant in SMARCAL1, negatively impacting SMARCAL1 function and exacerbating genome instability with tumor acquisition of somatic second hits in genes known to permit ALT (SMARCAL1 LOH, ATRX inactivation), resulting in OS development.

Table Legends

Table 1 Significant cancer-gene associations from the discovery and replication cohorts

References

1. Zhang J, Walsh MF, Wu G, et al: Germline Mutations in Predisposition Genes in Pediatric Cancer. N Engl J Med 373:2336-2346, 2015

2. Grobner SN, Worst BC, Weischenfeldt J, et al: The landscape of genomic alterations across childhood cancers. Nature 555:321-327, 2018

3. Newman S, Nakitandwe J, Kesserwan CA, et al: Genomes for Kids: The Scope of Pathogenic Mutations in Pediatric Cancer Revealed by Comprehensive DNA and RNA Sequencing. Cancer Discov 11:3008-3027, 2021

4. Fiala EM, Jayakumaran G, Mauguen A, et al: Prospective pan-cancer germline testing using MSK-IMPACT informs clinical translation in 751 patients with pediatric solid tumors. Nature Cancer 2021 2:3 2, 2021-02-15

5. Wagener R, Taeubner J, Walter C, et al: Comprehensive germline- genomic and clinical profiling in 160 unselected children and adolescents with cancer. European Journal of Human Genetics 2021 29:8 29, 2021-04-12

6. Ripperger T, Bielack SS, Borkhardt A, et al: Childhood cancer predisposition syndromes-A concise review and recommendations by the Cancer Predisposition Working Group of the Society for Pediatric Oncology and Hematology. American Journal of Medical Genetics Part A 173, 2017/04/01

7. Blake A, Perrino MR, Morin CE, et al: Performance of Tumor Surveillance for Children With Cancer Predisposition. JAMA Oncology 10, 2024/08/01

8. Sharma R, Lewis S, Wlodarski MW: DNA Repair Syndromes and Cancer: Insights Into Genetics and Phenotype Patterns. Frontiers in Pediatrics 8, 2020

9. Qin N, Wang Z, Liu Q, et al: Pathogenic Germline Mutations in DNA Repair Genes in Combination With Cancer Treatment Exposures and Risk of Subsequent Neoplasms Among Long-Term Survivors of Childhood Cancer. Journal of Clinical Oncology 38:2728-2740, 2020

10. Downing JR, Wilson RK, Zhang J, et al: The Pediatric Cancer Genome Project. Nat Genet 44:619-22, 2012

11. Ma X, Liu Y, Liu Y, et al: Pan-cancer genome and transcriptome analyses of 1,699 paediatric leukaemias and solid tumours. Nature 555:371-376, 2018

12. Hudson MM, Ness KK, Nolan VG, et al: Prospective medical assessment of adults surviving childhood cancer: Study design, cohort characteristics, and feasibility of the St. Jude Lifetime Cohort Study. Pediatric Blood & Cancer 56, 2011/05/01

13. Consortium TGP, Abecasis GR, Auton A, et al: An integrated map of genetic variation from 1,092 human genomes. Nature 491:56-65, 2012

14. Raghavan NS, Brickman AM, Andrews H, et al: Whole-exome sequencing in 20,197 persons for rare variants in Alzheimer’s disease. Ann Clin Transl Neurol 5:832- 842, 2018

15. Robison LL, Armstrong GT, Boice JD, et al: The Childhood Cancer Survivor Study: a National Cancer Institute-supported resource for outcome and intervention research. J Clin Oncol 27:2308-18, 2009

16. Worst BC, van Tilburg CM, Balasubramanian GP, et al: Next-generation personalised medicine for high-risk paediatric cancer patients - The INFORM pilot study. Eur J Cancer 65:91-101, 2016

17. Karczewski KJ, Francioli LC, Tiao G, et al: The mutational constraint spectrum quantified from variation in 141,456 humans. Nature 581:434-443, 2020

18. Pinto EM, Chen X, Easton J, et al: Genomic landscape of pediatric adrenocortical tumors. Nature communications 6, 2015 Mar 6

19. de Andrade KC, Lee EE, Tookmanian EM, et al: The TP53 Database: transition from the International Agency for Research on Cancer to the US National Cancer Institute. Cell Death Differ 29:1071-1073, 2022

20. Pinto EM, Ribeiro RC, Kletter GB, et al: Inherited germline TP53 mutation encodes a protein with an aberrant C-terminal motif in a case of pediatric adrenocortical tumor. Fam Cancer 10:141-6, 2011

21. Kim J, Vaksman Z, Egolf LE, et al: Germline pathogenic variants in neuroblastoma patients are enriched in BARD1 and predict worse survival. J Natl Cancer Inst 116:149-159, 2024

22. Adamovich AI, Banerjee T, Wingo M, et al: Functional analysis of BARD1 missense variants in homology-directed repair and damage sensitivity. PLoS Genet 15:e1008049, 2019

23. Brady SW, Liu Y, Ma X, et al: Pan-neuroblastoma analysis reveals age- and signature-associated driver alterations. Nature Communications 2020 11:1 11, 2020-10-14

24. Grange LJ, Reynolds JJ, Ullah F, et al: Pathogenic variants in SLF2 and SMC5 cause segmented chromosomes and mosaic variegated hyperploidy. Nat Commun 13:6664, 2022

25. Singh VK, Rastogi A, Hu X, et al: Mutational signature SBS8 predominantly arises due to late replication errors in cancer. Commun Biol 3:421, 2020

26. Liu H, Xu C, Diplas BH, et al: Cancer-associated SMARCAL1 loss-of- function mutations promote alternative lengthening of telomeres and tumorigenesis in telomerase-negative glioblastoma cells. Neuro-Oncology 25, 2023/09/05

27. Chen C, Qin N, Wang M, et al: Cancer germline predisposing variants and late mortality from subsequent malignant neoplasms among long-term childhood cancer survivors: a report from the St Jude Lifetime Cohort and the Childhood Cancer Survivor Study. Lancet Oncol 24:1147-1156, 2023

28. Feuerbach L, Sieverling L, Deeg KI, et al: TelomereHunter - in silico estimation of telomere content and composition from cancer genomes. BMC Bioinformatics 2019 20:1 20, 2019-05-28

29. Ballinger ML, Pattnaik S, Mundra PA, et al: Heritable defects in telomere and mitotic function selectively predispose to sarcomas. Science 379, 2023-01-20

30. Akhavanfard S, Padmanabhan R, Yehia L, et al: Comprehensive germline genomic profiles of children, adolescents and young adults with solid tumors. Nature Communications 11, 2020

31. Lippner E LT, Salgado C, et al .: Schimke Immunoosseous Dysplasia, in Adam MP FJ, Mirzaa GM, et al. (ed): GeneReviews. GeneReviews, NLM, 2023

32. Mirabello L, Zhu B, Koster R, et al: Frequency of Pathogenic Germline Variants in Cancer-Susceptibility Genes in Patients With Osteosarcoma. JAMA Oncol 6:724-734, 2020

33. Brosnan-Cashman JA, Davis CM, Diplas BH, et al: SMARCAL1 loss and alternative lengthening of telomeres (ALT) are enriched in giant cell glioblastoma. Modern Pathology 34, 2021/10/01

34. O’Sullivan RJ, Greenberg RA: Mechanisms of Alternative Lengthening of Telomeres. Cold Spring Harbor Perspectives in Biology 17, 2025

35. Braithwaite EK, Prasad R, Shock DD, et al: DNA Polymerase A Mediates a Back-up Base Excision Repair Activity in Extracts of Mouse Embryonic Fibroblasts. Journal of Biological Chemistry 280, 2005/05/06

36. Daley JD, Olson AC, Bailey KM: Harnessing immunomodulation during DNA damage in Ewing sarcoma. Front Oncol 12:1048705, 2022

37. Cole S, Gianferante DM, Zhu B, et al: Osteosarcoma: A Surveillance, Epidemiology, and End Results program-based analysis from 1975 to 2017. Cancer 128:2107-2118, 2022

Figure 1

A

Discovery Cohorts

Cases (n=5,993)

· PCGP (n=687)

· G4K (n=254)

· RTCG (n=1263)

· SJLIFE (n=2603)

· TARGET (n=1186)

Controls (n=14,477)

· 1000 Genomes

Project (n=2630)

· ADSP (n=11847)

Variant Calling

GATK Joint genotyping (n= 20,470)

Joint VCF

VCF

Quality control and post hoc filtering Genetic ancestry and sex calling

Variant Filtering

diana banana

189 DNA Repair Genes

· Rare (AF <0.05%)

· Protein truncating

· Splicing

· Missense (REVEL >0.7)

Replication Cohorts

St. Jude CCSS (n=959)

German Childhood Cancer Registry (n=186)

INFORM (n=349)

ALL-NOS NHL

Hematologic

TALL

AML

BALL

ATRT

ACPG

CNS

EPD

LGG

HGG

ACC

LIC

Solid

MEL

GCT

EWS

RMS

WLM

NBL

100

200

300

400

500 1700 1800

Number of Germline Samples

ALL-NOS

NHL

TALL

Hematologic

AML

BALL

ATRT

ACPG

EPD

CNS

LGG

HGG

ACC

LIC

MEL

GCT

EWS

Solid

RMS

WLM

NBL

Germline Variants in DDR Genes (%)

5,993 Pediatric cancer samples

Passed QC

6,636,054

Frameshift, Nonsense, Missense, Splice

2,536,772

Rare (gnomAD AF <0.05%)

1,403,469

In DDR genes

10,368

REVEL >0.7

2,059

ACC

AML

BALL

EPD

EWS

HGG 6

LGG

NBL

NHL

TALL

WLM

ATM

BARD1

BRCA2

BRIP1

CUL4A

DDB1

FANCE

FBXL2

FEN1

HELQ

MLH1

MUS81

PARP2

PMS2

POLD1

POLL

POLM

RMI1

SMARCAL1

SMC5

SPIDR

SSBP1

TDG

TP53

XRCC3

XRCC4

-log10(FDR)

Figure 2

R110Pfs*39

c.375+1G>A

N311Kfs*26

c. 1101-2A>G

V157F

R196*

V218M

G266E

R273C

E285V

ACC

TP53

NM_000546

100

150

200

250

300

350

TAD

P53

P53 tetramer

HGG

R158H

V274Lfs*31

C242F

R248Q

R273C

R112*

L95Kfs*2

Y180*

L447V

L480S

R641*

NBL

BARD1

NM_000465

100

200

300

400

500

600

700

RING

ANK repeat

BRCT

Q469*

R487L

EWS

POLL

NM_013274

100

150

200

250

300

350

400

450

500

550

BRCT

Nucleotidyltransferase

E126Vfs*12

c.380+1G>C

C417*

N940Y

SMC5

NM_015110

100

200

300

400

500

600

700

800

900

1000

1100

P-loop NTPase

Coiled Coil

FRAMESHIFT

NONSENSE

SPLICE

MISSENSE

BARD1

POLL

SMC5

N940Y#

L447V

L480S#

R487L#

SMC5: wild-type

BARD1:p.L447V

[0 - 138]

NBL

BARD1:p.L480S

SMC5:c.380+1G>C

SMC5:p.C417*

SMC5:c.380+1G>C

[0 - 88]

SMC5:p.N940Y

25%

50%

75%

100%

SBS1

Age

SBS8- Late Replication Errors

SMC5 [NM_015110]

SBS5

SBS18- ROS Damage

Unassigned

72873449

72880297

72887145

72893994

Exon 1

Exon 2

Exon 3

Exon 4 Exon 5

Figure 3

SMARCAL1 NM_014140

R114Qfs*4 L139Efs*3

L397Rfs*40 R490C

Q653*

R820H

Discovery Cohort

☐

100

200

300

400

500

600

700

800

900

Replication Cohort

☒

L139Efs*3

F279S

c.863-2A>G

L397Rfs*40

c. 1335-2A>T

R563*

c.2070+2

F801V

A838T

E848*

G857R

F941Lfs*31

RBD

☐ HARP

☐ Helicase ☒ FRAMESHIFT

☐ NONSENSE

☒ SPLICE

☐ MISSENSE

Discovery

Replication CCSS

INFORM

GCCR

ClinVar Classification

Clin Var

gnomAD (v4.1) AF

Cancer Relapse

☐ P

☐ <0.001%

Tumor Availability

☐ LP

AlphaMissense Prediction

☐ VUS

☐ 0.001%-0.01%

☐ NA

☐ 0.1%-0.5%

gnomADv4.1 AF

Bi-allelic in Tumor

Relapse

☒ Bi-allelic

SMARCAL1

☐ Yes

Pathogenic Germline

MLH1

☐ No

☐ Mono-allelic

Variants

NF2

Tumor Availability

Variant Type

RB1

☐ Yes

☐ Missense

CDKN2A/B

☐ No

☐ Stopgain

☐ Frameshift

Pathogenic Somatic Variants

TP53

AlphaMissense

☐ Splicing

☐ Inframe InDel

BAP1

☐ Likey Benign

NF2

☐ Likey Pathogenic

☐ Deletion

ATRX

☐ Ambiguous

☐ Not Available

R490C

R820H#

2.5% Osteosarcoma Prevalence

Osteosarcoma Development

F801V#

Wild-type SMARCAL 1

Germline Variants in SMARCAL1

Somatic second hit in ALT genes (SMARCAL1, ATRX, etc.)

G857R#

A838T

Genomic Integrity

Genomic Instability

F279S#

Efficient DNA replication and repair

Replication stress High DNA repair demand

Replication fork breaks Double strand breaks High DNA damage

Table 1 Summary of significant cancer:gene associations from discovery and replication cohorts

Discovery analysis

Gene	Cancer	Frequency in Cases	Frequency in Controls	Cancer Risk
Gene	Cancer	Frequency in Cases	Frequency in Controls	OR [95% CI]	Plogistic	FDR logistic
TP53	ACC	10 in 27 (37%)	31 in 14477 (0.21%)	289.2 [120.7 to 673.5]	2.29E-37	8.42E-34
TP53	HGG	5 in 206 (2.4%)	31 in 14477 (0.21%)	11.1 [3.9 to 26]	2.36E-06	0.0022
BARD1	NBL	6 in 485 (1.2%)	32 in 14477 (0.22%)	6 [2.3 to 13.2]	0.0001	0.0341
POLL	EWS	2 in 117 (1.7%)	14 in 14477 (0.1%)	23.8 [4.6 to 81.1]	0.0001	0.0319
SMC5	MB	4 in 257 (1.6%)	12 in 14477 (0.08%)	21.6 [6.4 to 60.1]	2.70E-07	0.0005
SMARCAL1	OS	6 in 230 (2.6%)	60 in 14477 (0.41%)	6.3 [2.4 to 13.5]	6.81E-05	0.0250

Replication analysis

Gene	Cancer	Frequency in Cases	Frequency in gnomAD v2.1 (non-cancer)	Cancer Risk
Gene	Cancer	Frequency in Cases	Frequency in gnomAD v2.1 (non-cancer)	OR [95% CI]	P Fisher
POLL	EWS	1 in 209 (0.5%)	199 in 236207 (0.17%)	2.8 [0.1 - 16.1]	0.298
SMC5	MB	1 in 209 (0.5%)	74 in 225820 (0.07%)	7.3 [0.2-42.2]	0.130
SMARCAL1	OS	15 in 604 (2.5%)	409 in 236367 (0.35%)	7.3 [4.0-12.2]	7.88E-09

Quartz 4

Explorer

40463577

Comprehensive investigation of DNA damage repair genes in children with cancer identifies SMARCAL1 as novel osteosarcoma predisposition gene

Abstract

Background

Methods

Patient Cohorts

Variant Calling and Filtering for the Discovery cohort

Statistical Analysis

Results

Germline variants in DDR genes across tumor types

Significant germline associations in DDR genes

Replication of SMARCAL1 as a novel osteosarcoma predisposition gene

Discussion

Acknowledgements

Data Availability Statement

Figure Legends

Figure 1 Germline DDR gene variants across pediatric cancers.

Figure 2: Germline mutations in DDR genes and their putative functional impacts.

References

Figure 1

A

Table 1 Summary of significant cancer:gene associations from discovery and replication cohorts

Graph View

Table of Contents

Backlinks